utils.py

Table of Contents

1. Introduction
2. "import" statements
3. Functions
4. Checklist of features across Streamlit and LocalWorkflow

1. Introduction

This contains a set of utility functions that are necessary to be invoked across various modules in the app.

2. "import" statements

• thefuzz
• Used for calculating SimpleRatio
• pandas
• Used to create dataframe out of CSV file
• xml.etree
• For parsing the contents of the XML file

3. Functions and classes

csv_parser(filename, start_value=70, end_value=73)

• creates a dataframe using pandas
• range can be specified by setting df_condition to "1"

find_last_closing_tag_index(xml_string)

• finds the last closing tag '>' to strip the code block from the XML output generated

strip_code_block(string)

• Used in LocalWorkflow and with HuggingFacePipeline
• extracts the XML output from the entire chat response of the LLM
• This ensures that XML_operations such as similarity processing and checking for general validity can be conducted even when using the LocalWorkflow

XML_Similarity(XML_responses, text_value)

• takes two XML files, depending on the index numbers chosen by the user from the dataframe
• creates a dictionary of elements and attributes
• calculates the similarity between each element and attribute from each of the XML files, comparing both elements and attributes of each file with each other.
• Reason: In the absence of consistency amongst similarly generated XML tags across various iterations, they are compared so that user can change serialisation of the numbering if it occurs across various tags
• WIP: It may be necessary to allow user to view and select each XML portion as per their preference, or check throughout the document
• Currently implemented only in Streamlit GUI

4. Checklist of features across formats

(TBD = To Be Developed)

Feature	Streamlit	LocalWorkflow	LegalDocML	LegalRuleML
Dataframe generation	Yes	Yes	Yes	Yes
strip_code_block	No	Yes	Yes	Yes
XML_similarity testing	Yes	TBD	Yes	Yes